Introduction

One of the basic principles of health physics and radiation protection related to nuclear reactors is the control of radioactive material released into the environment due to natural disasters like earthquakes or due to other reasons such as human errors or technical malfunctions. The atmosphere is one of the main reservoirs of radioactive materials from a nuclear reactor accident. The International Atomic Energy Agency (IAEA) has published a review of principles and practical applications of atmospheric dispersion models in case radionuclides are released into the environment in IAEA-TECDOC-379 entitled “Atmospheric dispersion models for application in relation to radionuclide releases” (IAEA 1986). Nuclear installations give rise to radionuclide release, albeit in small quantities under controlled conditions, but they also have the potential to release large quantities of radionuclides in the case of accidents (IAEA 1986). For these reasons, assessment of radiation doses and associated radiological risks for individuals and populations is required, in particular when the releases exceed the routine releases predicted at the design and licensing stage (IAEA 1986).

Calculation of annual total effective dose equivalent (TEDE) received by the members of the public around a nuclear reactor site is very important as far as human health and safety are concerned. The objective of this study is the simulation of atmospheric dispersion and the calculation of the TEDE received by residents of Tehran and parts of its neighboring provinces, due to a hypothetical accident in Tehran Research Reactor (TRR). During an incident with an uncontrolled source of radiation, protection of the public from unnecessary radiation exposure may require some interventions that may disrupt daily life of those living close to the site of radioactive release. Such interventions are termed protective actions. The radiological dose assessment for members of the public during a TRR radiological accident will determine if protective actions are needed for avoiding or minimizing radiation doses and thus protecting people’s health.

Simulations for hypothetical accidents have been carried out on local- and meso-scales. For example, a Gaussian plume model was used for radiological dose assessment, for a hypothetical severe accident of the TRR on a local scale of about 10 km around the site (Ahangari et al. 2017). A similar study was also performed by Raza and Iqbal for the Pakistan Research Reactor-1 (Raza et al. 2005). Baklanov estimated the possible impacts across Europe after a hypothetical accident at the Kola Nuclear Power Plant (NPP) using an isentropic trajectory model for forwarding trajectories calculation and Lagrange random displacement models for radionuclide transport simulation (Baklanov et al. 2002). According to these studies, gas and particulate radionuclides may be released into the atmosphere through the reactor stacks, and transported by the wind to populated zones in the reactor surroundings. Consequently, the TEDE needs to be assessed, to determine the potential health risk to the affected population and to determine which kind of protective actions is needed under accident conditions. Radionuclide effluent dispersion models are different in principle. The horizontal scale of different transport and diffusion processes are important issues for choosing an appropriate transport model (Zali et al. 2017; Rentai 2011). Instantaneous and short-term releases are frequently viewed as puff releases. A puff release model assumes that the release time and sampling times are very short compared to the travel time of the radioactivity from the source to the receptor. In contrast, continuous releases are generally modeled as a plume. The assumption in a plume model is that the release time is much greater that the travel time of radionuclides from the source to the receptor. Therefore, a plume model is suitable for distances up to 10 km, while a puff model is useful for distances of more than 30 km (Hanna et al. 1982, IAEA 1994). The terrain and meteorological data are further important issues in choosing a model.

Accordingly, in this work, simulation of long-range transport of radionuclides and dose assessment for residents of Tehran city province for a hypothetical TRR accident were performed by means of the HYSPLIT 4 code (Hybrid Single-Particle Lagrangian Integrated Trajectory model Version 4). This code considers terrain and meteorological data and is suitable for simulation of long-range airborne transport and deposition of radionuclides. The code applies a Lagrangian dispersion model for calculating trajectories including complex dispersion and deposition simulations. The HYSPLIT version 4 is a complete system for computing simple trajectories as well as complex dispersion and deposition patterns, and for using either puff or particle approaches (Draxler et al. 1998).

Materials and methods

TRR site description

The TRR is a 5 megawatt-thermal (MWth) pool-type light water research reactor, which became operational in 1967. Light water is used as moderator, coolant and shielding. Tehran as the capital and the most populated city in Iran is located in a highly seismic zone. Tehran area is located at the southern border of the central Alborz in north bounded by several active seismic faults. These faults are Mosha fault, north of Tehran fault and Eyvanki fault. South of Tehran fault is another main fault including north and south faults of Ray and Kahrizak. Figure 1 shows the map of Tehran faults (Arian and Bagha 2012).

Fig. 1
figure 1

Position of faults in Tehran and its surroundings

Full size image

Scenarios and radioactivity source term

In the assumed hypothetical severe accident, the reactor core melts as result of an earthquake. As a result, a fraction of radioactive substances is released from the TRR stack into the environment. The amount of radionuclides released into the environment is the radioactivity source term. The core inventory of radionuclides at the time of the accident was calculated with the ORIGEN-2 computer code. For these calculations, the TRR was assumed to have been operated continuously at its full power (5 MWth) and up to an average burnup of 40 GWd/t U. The cross-section library for the TRR core inventory calculations was generated by the Monte Carlo transport code MCNP. The resulting TRR total core activity included 1.11 × 1016 Bq for activation products, 8.80 × 1017 Bq for fission products, and 2.96 × 1016 Bq for actinides.

It was further assumed that radioactive substances were released from the TRR stack into the environment. To estimate the radionuclides released into the atmosphere, the following values for the release fractions have been considered taking into account filter efficiencies of the emergency ventilation system as given in the safety analysis report (AEOI 2009): The release fractions from the reactor core to the reactor containment were 10% for iodine, 100% for the noble gases, and 0.5% for other fission products. Also, the release fractions from the reactor containment to the atmosphere were 2% for iodine, 100% for noble gases, and 1.5% for other fission products. Based on these release fractions, Table 1 shows the activity of radionuclides released into the atmosphere.

Table 1 Radionuclides released into the environment (source term)
Full size table

Wind data

The wind direction and frequency of the wind speed as measured at the Tehran nuclear research metrological center were taken from the TRR safety analysis report. These data are presented in Table 2. According to this report, the dominant wind direction at the TRR site is from southwest to northeast and the predominant wind speed is 9–15 m/s, at 45 m height (AEOI, 2009).

Table 2 Percentage of time that the wind comes from a given direction and the wind speed as measured at a height of 45 m on the meteorological tower of the Tehran nuclear research center (hourly average over a period of ten years)
Full size table

Effective stack height

The effective height at which the radionuclides are released is not equal to the physical stack height, but can be higher due to the interplay between the stack release rate and the temperature difference between stack effluent and surrounding air. Therefore, in the calculation of the release of radioactive material from the stack, the effective stack height must be considered. According to Holland’s formula, the effective stack high is (Eq. 1):

$$H=h+\Delta H$$
(1)

where H is the effective stack height (m), h is the height of physical stack (m), \(\Delta\)H is the plume rise. The plume rise is given by Eq. 2:

$$\Delta H={v_{{\text{S}}~}}\left[ {1.5+\left( {2.68 \times {{10}^{ - 2}}P\left[ {\frac{{{T_{\text{S}}} - {T_{\text{a}}}}}{{{T_{\text{a}}}}}} \right]d} \right)} \right]$$
(2)

where \({v_{{\text{S~}}}}\) is the stack velocity (m/s), d is the stack diameter (m), P is the pressure (kPa), Ts is the stack temperature (ºK), and Ta is the air temperature (ºK).

Details on the specification of the TRR stacks required as input for the HYSPLIT code are presented in Table 3 (AEOI 2009).

Table 3 Specification of the TRR stacks
Full size table

The effective stack height depends on atmospheric stability, which in turn depends on the change of temperature with height and wind speed. Stability classes are defined for different meteorological conditions, which are dependent on wind speed and solar insolation during the day and cloud cover during the night. Atmospheric turbulence is categorized into six stability classes named A, B, C, D, E and F, with class A being the most unstable or most turbulent class, and class F the most stable or least turbulent class. For classes A, B, C and D, the effective stack height is 90 meters, while the effective stack heights are 75.9 and 81.7 meters for classes E and F, respectively.

Height of receptor

The height of receptor above ground at which TEDE values were calculated is 1.5 m (corresponding to a typical height at which people inhale air). Height of receptor has an important effect on TEDE calculations. The breathing rate is taken to be 2.66 × 10−4 m3 s−1 for an average human being (ICRP 23, 1975; Paquet et al. 2017).

Population distribution

Tehran is the most populated city in Iran, with about 10 million inhabitants. The TRR is situated in the northwest of the city. About 80 percent of Tehran city area is located within a radius of 10 km from the TRR site. There is no residential area closer to 500 meters from the reactor site. The distribution of the population within 16 city sectors and for different distances from the TRR site is presented in Figs. 2 and 3 (AEOI 2009).

Fig. 2
figure 2

Population within 5 km around the Tehran Research Reactor (TRR)

Full size image
Fig. 3
figure 3

Population around the Tehran Research Reactor (TRR) as a function of distance

Full size image

Metrological condition

The National Center for Environmental Protection (NCEP) of the United States of America performs a set of computer analyses and forecast operations. The Air Resources Laboratory (ARL) of the National Oceanic and Atmospheric Administration (NOAA) uses the NCEP model for dispersion modeling calculations and simulation of transport of air samples. Since 1989, ARL has gathered meteorological data for research studies. One of the environmental data assimilation systems used is the Global Data Assimilation System (GDAS) which is also used in the present study (Draxler et al. 1998). GDAS meteorological data for several consecutive years from 2005 to 2014 is freely available at the following URL (ftp://arlftp.arlhq.noaa.gov/pub/archives/gdas1). The meteorological data of the TRR site and its vicinity are listed in Table 4 (AEOI 2009).

Table 4 Meteorological data at the TRR site and in its vicinity (Atomic Energy Organization of Iran 2009)
Full size table

Software used (HYSPLIT)

The HYSPLIT model has been developed by ARL. The HYSPLIT version 4 model is a complete model for computing simple trajectories involving complex dispersion and deposition simulations using either puff or particle approaches (Draxler et al. 1998). The model consists of a modular library structure with main programs for each primary application: trajectories and air concentrations. Gridded meteorological data on a latitude–longitude grid or on one of the three conformal map projections (Polar, Lambert and Mercator) are required at regular time intervals. The dispersion of a pollutant is calculated using either the puff approach, the dispersal of a fixed number of particles, or a combination of both methods by assuming a puff distribution in the horizontal and particle dispersion in the vertical direction. The dispersion rate is calculated from the vertical diffusivity profile, wind shear, and horizontal deformation of the wind field. Air concentrations are calculated at a specific grid point for puffs and as cell average concentrations for particles (Draxler et al. 1998). Some essential data needed for these calculations are listed below:

Dry deposition

Dry deposition refers to the deposition of pollutants as they settle out from the atmosphere or are absorbed by plant tissues. Calculation of dry deposition is one of the key concepts and very important in atmospheric modeling. This is because dry deposition depends on land surface conditions and metrological conditions. Dry deposition for urban areas is quantified by values reported in (Baklanov and Sorensen 2001).

Wet deposition

Wet deposition refers to the transfer of pollutants from the atmosphere to the earth involving precipitation. Wet deposition is described separately by in-cloud and below-cloud processes. The required values of the efficiency of in-cloud transfer and of the rate of below-cloud absorption were taken from studies conducted by the Air Resources Laboratory of NOAA (Draxler and Hess 2004). The in-cloud process involves aerosol particles getting into cloud droplets or cloud ice crystals by acting as condensation nuclei, or being captured through collision. They can reach the ground surface when rain or snow forms within clouds. The below-cloud process happens when falling rain droplets or snow particles collide with aerosol particles through Brownian diffusion, interception, impaction and turbulent diffusion. In-cloud transfer is described by the concentration of the pollutant in rain (g/liter). In contrast, below-cloud absorption is described by an absorption time constant. The numerical values used for in-cloud transfer and below-cloud absorption are 4 × 10+4 (g/lit) and 5 × 10−5 (s−1), respectively (Draxler and Hess 2004).

Henry’s constant

Most studies revealed that noble gases do not show any dry or wet deposition on the ground surface. Instead, they are just transferred to the atmosphere through the processes of radioactive decay (Baklanov and Sorensen 2001; Draxler and Hess 2004). Henry’s constant for noble gases was considered equal to zero. This means that any health effects from radioactive noble gases in the contaminated plume (cloud submersion) are either due to inhalation or external exposure.

Resuspension of deposited particles

The rate of resuspension considered in this study is assumed to be \(1 \times {10^{ - 6}}\) which is proposed as a value in the model (Draxler and Hess 2004).

Density, shape and particle diameter

Density, shape and particle diameter are used to define a pollutant that is subject to gravitational settling and wet removal. A value of 0 for any of these parameters means that the pollutant is treated as a gas. In contrast, a value of 1 may be assigned to define the pollutant as a particle. These values are flags set in the HYSPLIT computer code. According to Kinser’s and Draxler’s approaches, the diameter for inhaled particles is considered to be 1 µm and zero for vapors and gases, respectively. (Draxler et al. 1998; Kinser 2001).

Data selection

Using the MATLAB software, 360 days were randomly chosen where the hypothetical accident might have happened (one day per week for a total study period of about 7.5 years). Then, the HYSPLIT model was run for each considered day. By adding the resulting radionuclide concentrations in air, some days such as 09.01.2007, 19.07.2008, 15.05.2009 and 05.11.2010 were identified as the worst days in each season over a period of seven and a half years. This is because the concentrations of the radionuclides in the air and deposited on the ground surface depend on parameters such as the amount of the released radionuclide, wind speed, atmospheric stability and so forth, which differed among the chosen days. It is noted that to achieve more precise results than obtained in the present study, increased computer run time is needed to cover the entire time period on a day-to-day basis, which was not possible in the present study. More specifically, if the program was run for every day of the year, 2555 runs would have been needed to cover the 7 years study period. Moreover, if one had four simulations per day at 00:00 am, 6:00 am, 12:00 pm, and 18:00 pm, in total about 10,220 runs would have been needed. Assuming that the minimum time required for each run is about 15 min, a total of 153,300 min would have been needed which is equivalent to 107 days. Therefore, and because the weather data files are given as weekly files, one day of each week as an example of that week was considered sufficient, in the present study. Thus, using the MATLAB program 4 days per month were selected as sample days for eight consecutive years from 2005 to 2012. Hence, 360 days were selected over the 7.5 years study period.

Overall procedure

The simulation of the radionuclides transport and dispersion from the source (reactor stack) was performed using the HYSPLIT model (HYSPLIT—Draxler and Hess 1998). The TEDE for adults was then calculated including deposition and concentration of the considered radionuclides, and also using the dose conversion factors (DCFs) recommended by the International Commission on Radiological Protection (ICRP) and the Federal Guidance Report (FGR) (Eckerman and Ryman 1993; ICRP 1975). Surfer 8 software was used to draw the shape of the dose distribution at different distances around the reactor. All calculations were performed for a hypothetical accident that occurred after an earthquake.

Results and discussion

Concentration and deposition of radionuclides

In case of an accident, the radionuclide concentrations in air and on the ground (due to radionuclide deposition) contribute to the total dose received by the public. Figures 4 show the distribution of the radionuclides concentration (Bqm−3) at the height of 1.5 m above the ground level for the postulated worst days in a hypothetical accident scenario. Figure 4 shows that the northeastern and eastern areas of Tehran province would have been contaminated on 09.01.2007, while the western areas would have been contaminated on 19.07.2008, including the densely populated areas of Tehran such as Eslamshahr, Robat Karim, Qods and Shahriar. As is depicted in Fig. 4, during these days a large area of Tehran province and Shemiran would have been contaminated by radioactivity. Accordingly, the Southeast of Tehran province would have been heavily contaminated on 05.11.2010.

Fig. 4
figure 4

Hypothetical concentration of radionuclides (Bqm−3) at a height of 1.5 m above ground level on 09.01.2007, 19.07.2008, 15.05.2009 and 05.11.2010

Full size image
Fig. 5
figure 5

Deposition of radionuclides (Bq m−2) one day after the hypothetical accident on a 09.01.2007, b 19.07.2008, c 15.05.2009 and d 05.11.2010

Full size image

Figure 5 llustrates the deposition of the radionuclides (Bqm−2) one day after the hypothetical accident on the assumed worst days. For example, as is shown qualitatively in Fig. 5a, the deposition of radionuclides at a distance of 1 km from the site of release in the northeast direction is 5.7 × 104 Bqm−2 and increases to 3.4 × 105 Bqm−2 at 2 km and then at 40 km reduces to 376 Bqm−2. Generally, the estimated deposition decreases with the distance from the site of release similar to the radionuclide concentration.

Fig. 6
figure 6

Distribution of hypothetical total effective dose equivalent (TEDE) on a 09.01.2007, b 19.07.2008, c 15.05.2009 and d 05.11.2010

Full size image

Total effective dose equivalent (TEDE)

The TEDE for adults has been calculated from some pathways such as cloud submersion, ground shine and inhalation, based on the FGR 12 database and ICRP Publication 25.

To calculate the dose from inhalation, the DCF was multiplied with the time-integrated air concentration; to calculate the dose from ground shine, the DCF was multiplied with the activity concentration on the ground. Then, TEDE was calculated by summing these values. Surfer 8 software was used to draw the shape of the dose distribution at different distances about 100 km around the reactor. Figure 6 shows the calculated dose distribution, every 0.01° in the horizontal direction and in the vertical direction, which is equal to 1.1 km and 0.9 km, respectively. For the hypothetical accident investigated in the present study, it has been assumed that exposure pathways are air submersion, ground shine and inhalation. Figure 6 shows the TEDE distribution resulting from the dispersion of radionuclides around the TRR reactor on the identified worst days. According to the obtained values, the received dose decreases with distance from the reactor. Figure 6 shows that the maximum TEDE dose received by a resident is 1700 µSv at a distance of 1 to 2 km from the site of release on 09.01.2007. Moreover, Fig. 6c shows that the maximum received dose by a resident is 1,200 µSv around the reactor site on 15.05.2009. The maximum TEDE and its distance from the reactor in the specified worst days are presented in Table 5. The TEDE in the hypothetical accident scenario for the populated cities of Tehran is given in Table 6. The maximum TEDE value was calculated to be 1,011 µSv in Nasimshahr which is located in the southwest of Tehran province. Interestingly, no TEDE dose was obtained for Damavand and Firuzkuh located in the Alborz range of mountains, acting as a barrier against the transport of radionuclides.

Table 5 Hypothetical maximum annual total effective dose equivalent (TEDE) and its distance from the TRR reactor
Full size table
Table 6 Hypothetical annual total effective dose equivalent (TEDE) for populated cities of Tehran province
Full size table

In principal, protective actions for people include evacuation and sheltering-in-place. Sheltering-in-place may be warranted in situations where evacuation poses a greater risk of exposure or physical harm. These protective actions would be taken if the TEDE doses are expected to exceed the dose limits presented in IAEA-GSR Part 7 document (IAEA-GSR Part 7, 2015). The hypothetical TEDE doses calculated in the present study for residents of Tehran province at various distances from the TRR rector site are less than the corresponding protective action dose limits. Accordingly, for the TRR accident condition investigated in the present study, no protective actions need to be implemented for residents of Tehran province.

To compare the dose values obtained in the present work with those of other similar studies, long-range atmospheric transport simulations are required which are not available for the hypothetical TRR accident discussed here. Comparison of dose values obtained using the Lagrangian dispersion model in HYSPLIT with those using a Gaussian plume model which is used in the other studies, is not standard and reasonable.

Conclusions

Simulation of radionuclide dispersion and radiological dose assessment for residents of Tehran province assuming a hypothetical TRR accident was carried out. Simulation of radionuclide dispersion was performed using the HYSPLIT4 model, which is appropriate for long-range airborne transport and deposition of radionuclides. Several parameters were considered in the simulations that affect the dose to members of the public around a nuclear reactor in the case of accidental release. The results obtained imply that maximum TEDE values for adult members of the public around the TRR site are less than the dose limits proposed for protective actions. Consequently, no protective action would be required.Footnote 1